Japanese Keyword Hack: Find and Remove the Spam Pages
A Japanese keyword hack plants thousands of Japanese-language spam pages under your domain that show up in Google while your own site looks normal — the pages are cloaked, served mainly to search engines. Remove the injected content and the way in, hide the URLs while Google recrawls, then clear any flags.
TL;DR: The spam pages are built to be invisible to you. Google's own guide describes the hack as cloaking: autogenerated Japanese pages in randomly named directories, pushing affiliate links to fake brand merchandise, while visitors see your normal site or a 404 (Google). Confirm it in the Security issues report and with
site:searches; clean the injected rows, files and users; then use Search Console's removals tool to hide the URLs while Google recrawls — a temporary hide of about six months, not a deindex. Permanent removal happens when Google recrawls the cleaned site and finds 404s. Finally, remove the attacker's Search Console access and verification token.
What is a Japanese keyword hack?
The symptom is the story: thousands of Japanese-language pages appear in Google under your domain — usually fake luxury-goods and brand listings — while the site you browse looks completely normal. Google's guide describes autogenerated Japanese text in randomly generated directories, for example example.com/ltjmnjp/341.html, monetized through affiliate links to fake brand merchandise (Google).
The mechanism is cloaking. The pages are served mainly to search engines, so regular visitors see your site as usual — or, just as confusingly, a 404. Google's guide puts it bluntly: "Don't be fooled! Hackers will try to trick you into thinking the page is gone or fixed when it's still hacked" (Google).
WordPress is the classic case — the injection lives in the database and asks nothing of your dashboard, which is why it goes unnoticed — but do not take the platform for granted. The same symptom can appear on other stacks, through different mechanisms such as server rewrite rules. What they share is the end state: spam pages indexed under your domain that keep growing until the injection is removed.
How to confirm it (signs you can check)
A few checks, five minutes:
- A
site:search. Search Google forsite:yourdomain.comand page through the results for URLs you never published — odd directories, numeric filenames, another language. If Google shows nothing, repeat the search on another search engine; Google's guide suggests exactly that. - A keyword search. Search for your domain together with a spam phrase. The categories vary, but the pattern is always content you did not write.
- The Security issues report. Open Search Console → Security & Manual Actions → Security issues; Google lists hacked-content problems it has found, with sample URLs (Google).
- The URL Inspection tool. Enter a suspect URL and inspect what Google sees — the tool "lets you see the underlying hidden content", which is how you catch cloaking that your own browser cannot see (Google).
What not to trust: your wp-admin. Everything can look normal in the dashboard while the injected rows sit in the database and the spam keeps being served. One more useful tell — if you get a verification notification you did not trigger, treat it as a strong signal that someone else has owner access (Google).
How the spam pages get in
The two common entry points are a vulnerable plugin or theme, and a hidden administrator account. From there the injection is layered, and the layers matter for cleanup:
- Database rows. Spam posts and pages in
wp_posts; payloads tucked intowp_options, often encoded; sometimeswp_postmetaas well (Rank Math). - Files. Backdoor PHP — attackers favour familiar names like
index.php,wp-load.php,404.phpandview.php— often wrapped in obfuscation such asbase64_decode,evalorgzinflate(Google). - Rewrite rules.
.htaccessrules that serve the spam only to Googlebot and Bingbot, or that dynamically fake an HTML verification file (Rank Math). - Cron and users. Scheduled tasks that regenerate content, and admin users you did not create.
Keep the details aligned with the cleanup below. If the spam reappears after a deletion, the entry point is still open — a pattern our guide to WordPress malware that keeps coming back unpacks.
Remove the injection: the cleanup order
Back up the files and the database first — even an infected copy. If you can, take the site offline or block new writes (maintenance mode, a host-level lock) while you work. Then follow this order, because backdoor PHP files and cron jobs are what regenerate the spam: clean the database first and they write it straight back.
- Backdoor files. Reinstall core, themes and plugins from official sources rather than trusting what is on disk, then hunt leftovers in uploads and theme files, checking the backdoor names above.
- Rewrite rules. Replace
.htaccesswith a clean copy unless you have custom rules worth preserving — a hidden verification-token rewrite dies with it. - Cron jobs and users. Remove scheduled jobs that regenerate content. Remove admin accounts you did not create through the WordPress Users screen; editing
wp_usersin SQL is only a fallback, and only after a backup. - Database. Only now delete the spam rows from
wp_posts, and inspectwp_optionsandwp_postmetafor injected payloads. - Credentials. Rotate every password and key: WordPress admins, hosting, SFTP, the database, and the authentication salts that force all sessions out. The WordPress malware removal pillar walks through each step, and the plugin-based alternative if you would rather not touch these files by hand.
- Verify. Fetch the spam URLs again: they should return 404 or 410 and serve nothing hidden. Check again a day later that no new spam rows, files or cron jobs have appeared — if they have, something is still regenerating them.
Hide the spam URLs while Google recrawls
Deleting the pages is not the end of it: they return 404, and Google removes them from the index only when it recrawls and sees that. On an infection with thousands of URLs, that can take weeks, and stubborn cases linger longer. The Search Console removals tool is how you get them out of sight meanwhile:
- Use a prefix request. Search Console → Removals → New request → Temporary removals, choosing "Remove all URLs with this prefix" hides every URL that starts with the same string — one submission can cover a whole spam directory.
- It hides fast, and temporarily. "The request usually takes up to a day to process", and a successful request lasts about six months (Google). John Mueller is explicit that this is a display switch, not an index removal: "This doesn't remove them from the index, but it hides them within a day" (Search Engine Journal).
- Permanent removal comes from the crawl. Serve 404 or 410 and let Google recrawl; a 404 is the right response code, and 410 says the URL is gone for good (Search Engine Journal). If a URL already returns 404 when you file the request, Google "will assume that the page is gone, and your block request will expire" — which is fine once the page truly is gone (Google).
- Do not block the paths in robots.txt. Blocking stops Google from recrawling them, and a page that cannot be crawled cannot be dropped from the index. Hide with the removals tool, and serve 404 or 410 underneath.
Remove the attacker's Search Console access
Do this while you are in Search Console for the removals, because the attacker's access outlives the spam pages. Google's guide notes that "the hacker will typically add themselves as a property owner" — an unexpected verification notification is the tell (Google).
- Open Settings → Users and permissions and remove any owner or user you did not add (Google). Make sure your own access is verified before you remove anyone else's.
- Delete the unused ownership tokens behind their access. A removed owner can re-verify while a token still exists, and tokens are not removed automatically (Google).
- Find the tokens everywhere they hide: an HTML file at your root, a meta tag, a DNS record — or a
.htaccessrule faking a verification file. If you cannot find an HTML token, check.htaccessfor that rewrite rule, and confirm it is dealt with by visiting a simulated token URL; it should return 404 (Google).
Why the site is still flagged in search
A "This site may be hacked" label under your listing is Google's own signal, and it goes away only after the site is clean and you request a review in the Security issues report; Google's help says a review can take from a few days to a few weeks. Our guide to removing "This site may be hacked" walks through the review request.
Other security vendors are separate. While the spam pages were live, crawlers and blocklist feeds recorded what your domain served, and those flags do not follow Google's review — each vendor owns its list, and each needs its own request. Check where you stand before assuming: our free blocklist check reads the major security vendors and shows which currently flag the domain. It is a reputation check, not a malware scan — it cannot tell you the injection is gone, only who is blocking you.
Still flagged after cleanup? The removal requests
Each vendor's list is its own. Clearing one does not clear the rest, and none of them clears on its own — so after the cleanup, and while the deindexing works its way through, the remaining work is a round of removal requests.
That is the step unflagdomain.com handles. For a one-time €49 per domain, we email a removal request to every flagging vendor that takes email — plain text, a separate request written for each vendor, with your own address as Reply-To — and prepare ready-to-paste text for the ones that only take a form, Google Safe Browsing's Search Console review included. A correctly formatted removal request is sent to every flagging vendor that accepts one by email, within 48 hours of payment (in practice within the first hour). The guarantee is the dispatch, never the delisting: each vendor decides on its own schedule. Automatic refund when the post-payment re-scan comes back with zero flags. And the boundary again: we do not scan for or clean malware — the work above comes first. Our catalog covers 136 active vendors.
The spam pages are the visible half
A Japanese keyword hack looks like an SEO problem, but it is a live compromise: the pages are generated, not uploaded, and they keep coming until the injection and its entry point are gone. Remove the backdoor files and cron jobs first, then clean the database; rotate credentials; hide the URLs while Google recrawls. Then find out who else noticed — a free check across the security vendors shows the flags the spam left behind, and the removal requests are ours to send.
It is a site compromise, most often on WordPress, that plants thousands of spam pages — usually advertising fake luxury goods in Japanese — that show up in Google under your domain. The pages are cloaked, visible mainly to search engines, so your site looks completely normal when you browse it yourself.
No. Deleting the pages makes them return 404, and Google removes them from the index only when it recrawls and sees that — which can take weeks on a large infection. Search Console's removals tool can hide them from results within about a day while that happens, but it is a temporary hide, not a permanent deindex.
No — blocking the URLs stops Google from recrawling them, and a page that cannot be crawled cannot be removed from the index. Serve 404 or 410 instead, and use the removals tool to hide the URLs while the crawl catches up.
The spam is aimed at search engines and generally not shown to regular visitors, though redirects and hidden links are possible. Either way it is a live compromise that will keep growing — and any search warnings it causes need the site cleaned first.
Google and other vendors each decide on their own schedule — there is no date to promise. Our part is the dispatch: A correctly formatted removal request is sent to every flagging vendor that accepts one by email, within 48 hours of payment (in practice within the first hour).